U.S. patent number 5,121,222 [Application Number 07/739,796] was granted by the patent office on 1992-06-09 for method and apparatus for producing binary picture with detection and retention of plural binary picture blocks having a thin line pattern including an oblique line.
Invention is credited to Seiichiro Ejima, Toshiaki Endoh, Hisaharu Kato, Yasuhiro Yamazaki.
United States Patent |
5,121,222 |
Endoh , et al. |
June 9, 1992 |
Method and apparatus for producing binary picture with detection
and retention of plural binary picture blocks having a thin line
pattern including an oblique line
Abstract
Binary picture reducing method and apparatus are disclosed,
wherein an original picture is divided for a block of a pattern
matrix composed of a predetermined number of picture elements,
thereby extracting a plurality of binary picture blocks. When the
plurality of binary picture blocks each agree with a predetermined
thin line pattern, the binary picture block is detected as a thin
line block. The thin line block thus detected is retained on an
output picture at a position corresponding to a predetermined
reduction ratio.
Inventors: |
Endoh; Toshiaki (Tokyo-To,
JP), Kato; Hisaharu (Tokyo-To, JP), Ejima;
Seiichiro (Ohmiya-Shi, JP), Yamazaki; Yasuhiro
(Warabi-Shi, JP) |
Family
ID: |
15474536 |
Appl.
No.: |
07/739,796 |
Filed: |
August 1, 1991 |
Related U.S. Patent Documents
|
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
Issue Date |
|
|
531944 |
Jun 1, 1990 |
|
|
|
|
Foreign Application Priority Data
|
|
|
|
|
Jun 14, 1989 [JP] |
|
|
1-149411 |
|
Current U.S.
Class: |
358/451;
358/448 |
Current CPC
Class: |
G06T
3/4007 (20130101) |
Current International
Class: |
G06T
3/40 (20060101); H04N 001/393 () |
Field of
Search: |
;358/298,448,451,453,455,456,458,462,464,467,443 |
References Cited
[Referenced By]
U.S. Patent Documents
Primary Examiner: Fuller; Benjamin R.
Assistant Examiner: Rogers; Scott A.
Parent Case Text
This is a continuation of application Ser. No. 07/531,944, filed
June 1, 1990 now abandoned.
Claims
What we claim is:
1. A binary picture reducing method comprising:
a first step of dividing an original picture into blocks each
having a pattern matrix composed of a predetermined number of
picture elements for extracting a plurality of binary picture
blocks;
a second step wherein, when individual blocks of the plurality of
binary picture blocks each agree with a predetermined thin line
pattern including an oblique line, detecting each individual binary
picture block as a thin line block; and
a third step wherein each thin line block detected is retained on
an output picture at a position corresponding to a predetermined
reduction ratio.
2. A binary picture reducing apparatus comprising:
blocking means for dividing an original picture into a plurality of
blocks each of a pattern matrix composed of a predetermined number
of picture elements to extract the blocks as binary picture
blocks;
decision means for detecting, when the plurality of binary picture
blocks each agree with a predetermined thin line pattern including
an oblique line, each binary picture block as a thin line block;
and
retaining means for retaining each thin line block detected on an
output picture at a position corresponding to a predetermined
reduction ratio.
Description
BACKGROUND OF THE INVENTION
The present invention relates to processing for the production of a
reduced picture of a binary picture.
In present G4 facsimile 200, 240, 300 and 400 ppi (Pels/inch) are
used as standardized resolution for paper sizes of ISO A4, ISO B4
and ISO A3. Pictures of such various paper sizes and such various
resolutions are now employed in facsimile, and the conversion of
resolution (a picture reducing system) is requisite for the
intercommunication of pictures of different paper sizes and
resolutions. A variety of picture reducing systems have been
proposed. Typical systems are such as follows:
(1) SPC method (Gobo and Kirihara: One system for conversion of
facsimile line density, Processing of National Conference of the
Institute of Image Electronics Engineers of Japan, Vol. 7, No. 1,
(1978))
This is a system which uses that one of picture elements on the
original picture which is the closest to the picture element on a
reduced picture.
(2) Projection method (Morita, Komachi and Yasuda: System for
high-speed conversion of picture element density based on
projection method, The Journal of the Institute of Image Engineers
of Japan, Vol. 11, No. 2, (1982), for example) This is a system in
which the mean density of original picture elements projected onto
picture elements of a reduced picture is obtained and the value is
processed using a threshold value to obtain the value of each
picture element.
A system that has been proposed to retain thin lines is a TP method
(Wakabayashi, Kawanishi and Adachi: Reduction and conversion method
which prevents erasure of thin lines, The Transactions of the
Institute of Electronics, Information and Communication Engineers
of Japan, Vol. J70-D, No. 4 (1989. 7), for example), but this
method is directed to the retention of vertical and lateral lines
and does not take oblique lines into account. Thus, the
conventional picture reducing systems are poor in the retention of
thin line, and hence suffer from serious deterioration of picture
quality.
SUMMARY OF THE INVENTION
It is therefore an object of the present invention to provide a
binary picture reducing method and apparatus which permit retention
of thin lines including oblique lines with decreasing degredation
of the picture quality.
To attain the above objective, the binary picture reducing method
and apparatus of the present invention are constituted as
follows:
(1) A binary picture reducing method comprising:
a first step wherein an original picture is divided for a block of
a pattern matrix composed of a predetermined number of picture
elements, thereby extracting a plurality of binary picture
blocks;
a second step wherein, when the plurality of binary picture blocks
each agree with a predetermined thin line pattern, the binary
picture block is detected as a thin line block; and
a third step wherein the thin line block thus detected is retained
on an output picture at a position corresponding to a predetermined
reduction ratio.
(2) A binary picture reducing apparatus comprising:
blocking means whereby an original picture is divided for a block
of a pattern matrix composed of a predetermined number of picture
elements to thereby extract a plurality of binary picture
blocks;
decision means whereby, when the plurality of binary picture blocks
each agree with a predetermined thin line pattern, the binary
picture block is detected as a thin line block; and
retaining means whereby the thin line block thus detected is
retained on an output picture at a position corresponding to a
predetermined reduction ratio.
BRIEF DESCRIPTION OF THE DRAWINGS
The present invention will be described in detail below with
reference to accompanying drawings, in which:
FIG. 1 is a pattern diagram shown an example in which thin lines
disappear in the cases of the SPC method and the projection
method;
FIG. 2 shows examples of thin line decision patterns;
FIG. 3 is a pattern diagram showing the relationship between an
original picture and its reduced picture;
FIG. 4 is a pattern diagram showing the classification of thin line
retaining patterns;
FIG. 5 is a process flowchart for using this invention system and a
conventional system individually in dependence on whether the
process being performed is thin line processing or not;
FIG. 6 is a block diagram showing the picture reduction processing
according to the present invention;
FIG. 7 is a block diagram in the case of using the picture
reduction processing of the present invention and the conventional
picture reduction processing in dependence on whether the process
being performed is thin line processing or not; and
FIG. 8 is a diagram showing examples of additional thin line
retaining patterns.
DETAILED DESCRIPTION
In general, the SPC method and the projection method are poor in
the retention of thin lines. FIG. 1 shows an example in which thin
lines disappear. These defects can be eliminated in accordance with
the present invention.
The present invention will hereinafter be described in detail.
Whether a thin line is to be retained or not is determined
depending on whether or not is agrees with such a 3.times.3 mask
pattern as shown in FIG. 2. A white thin line pattern on a black
background can be checked by use of a black and white reversed
pattern of the 3.times.3 mask pattern depicted in FIG. 2. For the
sake of brevity, the following description will be given using only
the pattern shown in FIG. 2.
The relationship between an original picture and a reduced one is
defined as shown in FIG. 3.
R(M, N) (M, N=0, 1, 2, . . . ): original picture
R(M, N)=1: black picture element
R(M, N)=0: white picture element
P(i, j) (i, j=0, 1, 2, . . . ): reduced picture
P(i, j)=1: black picture element
P(i, j)=0: white picture element
Let it be assumed that the coordinates (M, N) are positioned at the
center of the picture element (N, N) and the coordinates (i, j) at
the center of the picture element P(i, j).
The picture element spacings of the original picture in vertical
and horizontal directions are normalized with 1. Since the
reduction ratios in the horizontal and vertical directions are
.alpha. and .beta., respectively, the picture element spacings of
the reduced picture in the horizontal and vertical directions are
.alpha. and .beta. on the original picture. Further, the
coordinates of the picture element P(i, j) on the reduced picture
are defined as x.sub.i, y.sub.i, and s.sub.i, t.sub.j are defined
as follows:
[]being Gauss' notation.
Algorithm 1
Features of algorithm 1 will hereinbelow be given.
(1) Letting the reduction ratios in the horizontal and vertical
directions be represented by .alpha. and .beta., respectively, the
following reductions can be effected:
(Mx and Nx being natural numbers, where Mx=1, . . . , Nx-1)
(My and Ny being natural numbers, where My=1, . . . , Ny-1)
(2) No thin lines are erased on the reduced picture.
(3) Picture elements of the original picture can sequentially be
processed following a left to right, top to bottom scanning
system.
Step 1:
Zeros are set in all picture elements on the reduced picture R(m,
n) (m, n=1, 2, 3, . . . ).
Step 2:
A 3.times.3 pattern composed of the picture element P(i, j) and
eight neighboring picture elements is obtained, and it is checked
whether or not this pattern agrees with the 3.times.3 mask pattern
shown in FIG. 2. If they agree, then set
Step 3:
The address of the next picture element is set by updating i,
j.
Step 4:
The process is finished after all picture elements are processed.
If not, the process returns to step 2.
With the algorithm 1, thin lines are not erased on the reduced
picture but may sometimes become thick.
Next, a description will be given of an algorithm which retains
thin lines intact on a reduced picture.
(Algorithm 2)
Features of the algorithm 2 are as follows:
(1) Letting the reduction ratios in the horizontal and vertical
directions be represented by .alpha. and .beta., respectively, the
following reduction is possible:
(Mx and Nx being natural numbers, where Mx=1, . . . , Nx-1)
(My and Ny being natural numbers, where My=1, . . . , Ny-1)
(2) No thin lines are erased on the reduced picture.
(3) Picture elements of the original picture can be sequentially
processed following the left to right, top to down scanning system
which is employed in facsimile. Thin line decision patterns are
classified into three types are shown in FIG. 4. With type A mainly
indicative of a horizontal line and type B mainly indicative of a
vertical line it is basically necessary to take into account the
retention of lines in the X or Y direction. As for the type C
indicative of an oblique line, the retention of lines in both of
the X and Y directions must be taken into consideration.
This algorithm will hereinbelow be described.
Step 1:
Zeros are set in all picture elements of the reduced picture R(m,
n) (n, n+1, 2, 3, . . . )
Step 2:
A 3.times.3 pattern composed of the picture element R(i, j) and
eight neighboring picture elements is obtained and it is checked
whether or not this pattern agrees with the 3.times.3 mask pattern
shown in FIG. 4. If they do not agree, then the process proceeds to
step 5.
Step 3:
If R(s.sub.i, t.sub.j)=1, the process proceeds to step 5.
Step 4:
The process follows the following procedure.
In the case of type A (A-1.about.A-5):
In the case of type A - 1, A - 5:
In the case of type B (B-1.about.B-5):
In the case of type B-1:
In the case of type B - 5:
In the case of type C - 1:
In the case of type C - 2:
Step 5:
i, j are updated and the address of the next picture element is
set.
Step 6:
The process is finished when all picture elements have been
processed. If not, then the process proceeds to step 2.
In this algorithm 2 the reduction ratio .alpha. in the horizontal
direction and the reduction ratio .beta. in the vertical direction
are limited as follows:
(Mx and Nx being natural numbers, where Mx=Nx/2, . . . , Nx-1)
(My and Ny being natural numbers, where My=Ny/2, . . . , Ny-1)
In this instance, if the picture element at the center of the
3.times.3 pattern which agrees with the 3.times.3 mask pattern
shown in FIG. 2 is referred to as a picture element forming a thin
line, and if a picture element train composed of picture elements
forming the thin line is defined as a thin line in the original
picture, it is retained as a thin line also in the reduced picture
through use of this algorithm 2.
Accordingly, the thin line is retained even if reduction processing
is repeated a plurality of times. Hence, a picture can be reduced
down to 1/2 or less by repeating the algorithm a plurality of
times.
To overcome the defects of the prior art systems, it is also
possible to combine them with this algorithm. In the case where
picture elements on the original picture agree with the thin line
decision pattern shown in FIG. 2, this algorithm is employed for
determining picture elements to be reduced, and in other cases, the
conventional systems are used. That is, in the case of the
algorithm 1, processing of step 2 is incorporated into the block
diagram of a thin line retaining algorithm shown in FIG. 7, and in
the case of the algorithm 2, processing of steps 2 to 4 is
incorporated.
Embodiment
Next, a detailed description will be given, with reference to the
accompanying drawings, of an embodiment of the present invention.
FIG. 6 is a block diagram illustrating the circuit arrangement for
performing the algorithm 2. Reference numerals 11, 12, 13, 14 and
15 indicate a memory for an original picture, a memory for a
reduced picture, an address set circuit, a 3.times.3 mask check
circuit, and a picture element value decision circuit,
respectively. At first, a picture to be processed is set in the
original picture memory 11. The address set circuit 13 manages
addresses of picture elements to be processed in the original
picture memory 11 and addresses of picture elements to be processed
in the reduced picture memory 12. The address set circuit 13
transfers to the 3.times.3 mask check circuit 14 the values of the
picture element to be processed in the original picture memory 11
and eight adjoining picture element. The 3.times.3 mask check
circuit 14 checks whether or not the picture element values agree
with the pattern shown in FIG. 2. If they do not agree, then the
mask check circuit instructs the address set circuit 13 to process
the next picture element. If they agree, then the mask check
circuit transfers to the picture element decision circuit 15 values
of the picture element to be processed and the eight neighboring
picture elements. The address set circuit 13 reads out of the
reduced picture memory 12 a picture element associated with the
picture element to be processed and transfers it to the picture
element value decision circuit 15. The picture element value
decision circuit 15 determines the reduced picture element value on
the basis of steps 3 and 4 of the algorithm 2. Where the reduced
picture value is determined to be 1 (i.e. a thin line retaining
picture element), the value is written into the reduced picture
memory 12 at an address received from the 3.times.3 mask check
circuit 14. After this, the picture element value decision circuit
15 instructs the address set circuit 13 to process the next picture
element. Where the reduced picture element value is 0 (i.e. not the
thin line retaining picture element), the picture element value
decision circuit 15 instructs the address set circuit 13 to process
the next picture element. While in FIG. 6 there is shown an example
of the circuit structure for performing the algorithm 2, a block
diagram of an example of the circuit structure for executing the
algorithm 1 can easily produced simply by modifying the contents of
processing of the picture element value decision circuit 15.
FIG. 7 illustrates in block form an example of the circuit
structure in the case where the algorithm 2 is combined with the
prior art system. Reference numerals 11, 12, 13, 14, 15 and 16
indicate an original picture memory, a reduced picture memory, an
address set circuit, a 3.times.3 mask check circuit, a picture
element value decision circuit, and a conventional picture reducing
circuit, respectively. The original picture memory 11, the reduced
picture memory 12, the address set circuit 13 and the picture
element value decision circuit 15 perform the same processing as
are executed by those in FIG. 6. The 3.times.3 mask check circuit
14 checks whether or not values of the picture element to be
processed and eight adjoining picture element agree with the
pattern shown in FIG. 2. If they agree, the values of the picture
element to be processed and the eight picture elements are
transferred to the picture element value decision circuit 15. If
they do not agree, the conventional picture reducing circuit 16 is
instructed to process the picture element. The conventional picture
element reducing circuit 16 reads out required picture element
values from the original picture memory 11 and the reduced picture
memory 12 and determines and writes the reduced picture element
value into the reduced picture memory 12.
Although the above description has been given using the 3.times.3
pattern, it is also possible to employ different patterns as long
as they permits the decision of a line element. For example, such
patterns as shown in FIG. 8 can be utilized in addition to the
patterns depicted in FIGS. 2 and 4. Moreover, the size of the
pattern matrix need not always be limited to 3.times.3 but may also
be 3.times.4, 4.times.3, 4.times.4, or 5.times.5, for instance.
While in the above the algorithms 1 and 2 have been described as a
simple and a detailed example of the thin line retaining algorithm
of the present invention, other algorithms can also be employed.
For instance, the conditions in step 4 of the algorithm 2 can be
modified as follows:
EXAMPLE 1
Step 4: The process follows the following procedure.
In the case of type A:
In the case of type B:
In the case of type C-1:
In the case of type C-2:
EXAMPLE 2
Step 4: The process follows the following procedure.
In the case of type A:
In the case of type B:
In the case of type C-1:
In the case of type C-2:
As described above in detail, the present invention prevents thin
lines from being erased during the reduction of a picture. Besides,
blurring of thin lines including oblique lines appreciably
decreases than in the prior art. Accordingly, the present invention
substantially improves degradation of the picture quality during
the reduction of a picture which are caused by the erasure and
blurring of thin lines in the prior art.
* * * * *